NO Need to Worry about Adversarial Examples in Object Detection in Autonomous Vehicles

perturbation of input may cause network to misclassify objects

• while not perceptible to human eyes
• generalize over diff. network architectures
• are they still adversarial in real world?(e.g. as an image)
  • yes…but in a short distance and fixed view
  • no…for most of the time

results

• distance: yes
• angle: yes…but not tested